Comparative Analysis of the Extracellular Matrix Proteome across the Myotendinous Junction

Abstract

The myotendinous junction is a highly interdigitated interface designed to transfer muscle-generated force to tendon. Understanding how this interface is formed and organized, as well as identifying tendon- and muscle-specific extracellular matrix (ECM), is critical for designing effective regenerative therapies to restore functionality to damaged muscle-tendon units. However, a comparative analysis of the ECM proteome across this interface has not been conducted. The goal of this study was to resolve the distribution of ECM proteins that are uniformly expressed as well as those specific to each of the muscle, tendon, and junction tissues. The soleus muscles from 5-month-old wild-type C57BL/6 mice were harvested and dissected into the central muscle (M) away from tendon, the junction between muscle and tendon (J) and the tendon (T). Tissues were processed by either homogenizing in guanidine hydrochloride or fractionating to isolate the ECM from more soluble intracellular components and then analyzed using liquid chromatography-tandem mass spectrometry. Overall, we found that both tissue processing methods generated similar ECM profiles. Many ECM were found across the muscle-tendon unit, including type I collagen and associated fibril-regulating proteins. The ECM identified exclusively in M were primarily related to the basal lamina, whereas those specific to T and J tissue included thrombospondins and other matricellular ECM. Type XXII collagen (COL22A1) was restricted to J, and we identified COL5A3 as a potential marker of the muscle-tendon interface. Immunohistochemical analysis of key proteins confirmed the restriction of some basal lamina proteins to M, tenascin-C to T, and COL22A1 to J. COL5A3, PRELP, and POSTN were visualized in the tissue surrounding the junction, suggesting that these proteins play a role in stabilizing the interface. This comparative map provides a guide for tissue-specific ECM that can facilitate the spatial visualization of M, J, and T tissues and inform musculoskeletal regenerative therapies.

Keywords: mass spectrometry; matrisome; mouse; muscle; tendon.

Figure 1.

Experimental workflow. (A) Regions of muscle (M), junction (J), and tendon (T) of 5 month-old murine soleus isolated for ECM proteomic analysis. Scale bar = 2 mm. (B,C) Steps for guanidine HCl (GuHCl; B) and tissue fractionation (C). See the Materials and Methods section for definition of abbreviations.

Figure 2.

Distribution of proteins identified by LC–MS/MS. (A) Raw intensities of GuHCl homogenates and CS and insoluble fractions from the fractionation protocol were plotted as percent of raw intensity and tissue compartment. Tissue was harvested from 5 month-old murine soleus muscle–tendon units. M = muscle; J = junction; T = tendon; average of n = 3 biological replicates. Two-way ANOVA revealed that the effect of cellular compartments between tissues was significant (p < 0.0001). For Tukey’s multiple comparisons post hoc analysis comparing the effect of tissue on % raw intensity of cellular compartments, see Table S2. (B) Raw intensities of the matrisome compartment from (A) were divided into different ECM categories (see Table S1). Two-way ANOVA revealed that the effect of ECM category was significant between tissues (p < 0.0001). For Tukey’s multiple comparisons post hoc analysis comparing the effect of tissue on % raw intensity of ECM category, see Table S5.

Figure 3.

Heat map comparing raw intensities of matrisome proteins identified in GuHCl homogenate from muscle (M), junction (J), and tendon (T) tissue. Raw intensities were log₁₀ transformed, and individual boxes represent each biological replicate for n = 3 M, T, and J samples (see also Table S1). Tissue was harvested from 5 month-old murine soleus muscle–tendon units. Rows were manually grouped to indicate the proteins found in specific tissues (see alternative arrangement in Figure S2).

Figure 4.

Immunolocalization of ECM to distinct regions of the soleus muscle–tendon unit. (A) TNC (green) was restricted to the tendon, whereas type XXII collagen (COL22A1, red) was found only at the J and where the T extended into the M (arrowheads; see also Figure S4). Nidogen-2 (NID2, blue) was localized to the M, consistent with the proteomics data (Table S1). (B) Type V collagen was found in all three tissues (COL5A1/A2/A3, green); however, the α3 chain of type V collagen (COL5A3, red) was visualized in the M, T, and J, contrasting with the proteomic identification of COL5A3 only in the J (Table S1). The α2 laminin chain (LAMA2, blue) was restricted to the J and basal lamina of M. (C) Prolargin (PRELP, red) was enriched in T and J tissues, overlapping with TNC (green) expression. (D) Periostin (POSTN, red) was only visualized at the interface between M (MY32, blue) and T (TNC, green) tissues. Images representative of N = 3 biological replicates; bars = 100 μm.

Figure 5.

Summary of ECM identified in both GuHCl and fractionation samples. Proteins in black were consistent between protocols, whereas proteins in gray were found in both data sets but with different tissue distributions. If a protein was found either in tendon alone or tendon + junction, it was included in the tendon group. Similarly, if a protein was found in either muscle alone or muscle + junction, it was included in the muscle group. COL5A3 is underlined to highlight the discrepancy between the proteomic and IHC data (see Figures 3, 4 and Table S1).